AI-Generated Celebrity New Year Videos Spark Personality Rights Violations
AI-generated deepfake videos of celebrities performing New Year greetings became popular on Chinese social media in early 2026, raising reported concerns about unauthorized use of celebrities' likenesses, voices, and copyrighted content such as Spring Festival Gala footage. Domestic AI platforms employed safeguards, but overseas and unregulated tools lacked equivalent protections, and social media templates spread widely with minimal legal warnings.
What happened
Around Chinese New Year 2026, AI-generated videos of celebrities offering New Year greetings went viral on Chinese social media platforms. Templates allowed users to upload their own photographs to appear alongside AI-synthesized versions of celebrities including Chen Peisi, Liu Dehua, Mbappé, and deceased performers such as Zhao Lirong and Feng Gong.
Some videos used material extracted from protected sources such as a 1989 Spring Festival sketch featuring Chen Peisi and Zhu Shizhao, with users applying deepfake technology to add their own facial features or create face-swapped versions of the performers. Other templates featured celebrity images performing specific choreography or delivering personalized blessings.
Reported impact
Legal experts identified multiple violations under Chinese civil law. Unauthorized use of a celebrity's likeness without consent violates portrait rights. Cloning a person's voice for synthetic videos raises separate infringement concerns under personality rights protections.
Templates incorporating footage from protected content such as the Spring Festival Gala could constitute copyright infringement, with statutory damages reaching up to 5 million yuan. Legal experts warned that synthetic content falsely portraying individuals could also constitute name and reputation rights violations if it involves insulting or false information.
Index analysis
Domestic Chinese AI platforms deployed strict controls to prevent unauthorized generation of celebrity likenesses, filtering by keyword, model restrictions, and user reports. However, some overseas platforms and unregulated tools lacked equivalent safeguards, allowing users to generate highly similar celebrity images through text prompts.
On social media, ease of use of the AI templates—requiring only photo upload and one-click sharing—enabled rapid viral spread among users, particularly older adults unfamiliar with the legal risks of personality rights. The templates included no platform-level warnings about intellectual property or personality rights infringement.
Governance controls
Legal experts recommended platforms implement pre-upload review mechanisms requiring authorization documents for templates containing celebrity images or voices, and embed visible AI-generation labels and invisible digital watermarks. Mid-stream controls should include AI content detection models, graduated penalties for repeat infringement accounts, and expedited response to rights holder complaints.
Platform responsibilities also include post-removal procedures such as blacklisting repeated infringement templates and publishing enforcement results. Platforms were advised to establish community standards explicitly prohibiting unauthorized use of personality rights, deploy risk warnings at template selection and pre-publishing stages in language accessible to older users, and conduct public education on copyright and personality rights protections.
Classification
Sources and evidence
Sources not yet published
This record was researched and written by the Index from original reporting. Its sources are still being verified by an editor, and they are listed here once that check is complete. Until then the record carries no reviewer.